Analysis of Personal Thermal State Using Machine Learning Algorithms to Prevent Heatstroke

Personal thermal state (PTS) refers to the individual’s thermal condition. PTS varies from person to person considering both internal and external factors. Internal factors such as a person’s physical fitness, body weight, body mass index (BMI), skin color, etc. can impact a person’s responses to any thermal changes. On the other hand, external factors such as physical activity, clothes, temperature, humidity, etc. can also influence someone’s thermal experience. PTS is a vital indicator that allows individuals to express their thermal conditions which may be comfortable or uncomfortable depending on a particular situation. PTS indicating ‘very hot’ with other health-related data may suggest the initial state of thermal stroke or heatstroke. Heatstroke is now a common and alarming syndrome due to global warming, which occurs due to the overheating condition of the human body. If neglected, heatstroke can create a life-threatening situation. This study proposes machine learning (ML) models to analyze PTS which may prevent heatstroke occurrences. We have developed models using a total of 9 ML algorithms where the Extra Tree (ET) algorithm provided an accuracy of 0.99858. On the contrary, K-Means Clustering provided very poor outcome on the particular dataset.